计算机工程与应用 ›› 2016, Vol. 52 ›› Issue (5): 179-184.

• 图形图像处理 • 上一篇    下一篇

改进的SLTP方法在掌纹识别中的应用

彭晏飞,周  南,林  森   

  1. 辽宁工程技术大学 电子与信息工程学院,辽宁 葫芦岛 125105
  • 出版日期:2016-03-01 发布日期:2016-03-17

Improved SLTP feature extraction algorithm for palmprint recognition

PENG Yanfei, ZHOU Nan, LIN Sen   

  1. School of Electronics and Information Engineering, Liaoning Technical University, Huludao, Liaoning 125105, China
  • Online:2016-03-01 Published:2016-03-17

摘要: 对掌纹识别进行研究,提出了基于改进的软直方图局部三值模式(SLTP)的掌纹识别方法。该方法先对掌纹训练样本进行能量函数的构造,然后用梯度下降法对能量函数进行优化,得到最佳的模糊隶属度函数,进而对掌纹的特征进行提取,最后用Chi概率统计的相似度度量方法进行匹配识别。在PolyU掌纹数据库和IITD掌纹数据库中进行实验验证,结果表明,在相同的训练样本下,改进的SLTP方法相比于SLTP方法,识别率得到提高。从而证明了该方法的有效性。

关键词: 掌纹识别, 软直方图局部三值模式, 梯度下降法, 模糊隶属度函数, Chi概率统计

Abstract: This paper presents a new method of palmprint identification which is based on improved SLTP to improve the effect of palmprint identification. Firstly, an energy function is proposed for the training samples. Then the gradient descent method is used to optimize the energy function and get fuzzy membership function. After this, it extracts the features of palmprint. Finally, a similarity measure method which is based on Chi probability and statistics is used to implement the matching recognition. The experimental results on PolyU palmprint database and IITD palmprint database show that compared with the SLTP method, the improved SLTP method which is proposed can effectively improve the recognition rate. The proposed method has better effect.

Key words: palmprint recognition, Soft histogram Local Ternary Patterns(SLTP), gradient descent, fuzzy membership function, Chi probability and statistics